imm

Associationsi from m tile Fossil Record

URBAN STONE DECAY .•---sc- Plant-Insect

• t-i-.i*'^ & Associations ^-''•:^ FROM THE Fossil Record v>i:::=^ r The coarsest characterization of life on Earth is probably one of teeming diversity. As terrestrial organisms ourselves, we are impressed by the enormous variety of -.^ life on land, and in particular, by the countless taxa of insects and vascular plants that constitute much of the écologie underpinning for terrestrial ecosystems. Any understanding of this diversity requires documentation of the associations among these two most pervasive groups of organisms. Historically, vascular plants and insects have provided the requisite numbers of species to support Ü\is ever-widening web of écologie interdependence. We know from the earhest fossils that only a few species of insects co-existed with primitive plants during the Early Devonian, principally as detritivores with monotonous diets of dead tissues, but also as a few nonspecialized herbivores. However, 400 miUion years later, after evolutionary radiations of land plants had produced complex ecosys- tems, approximately 280,000 species of vascular plants and several million to a few tens of millions of insects came to dominate the terrestrial world. Integral to this escalation of plant and insect numbers are the associations among these organisms, which • together with interacting fungi • represent a significant multiple of the individual • A diversities of the principal actors. 'i^^'îïy^- Consider a milkweed plant. Within a single photosynthesizing individual, a dozen species of herbivorous insect consumers can co-exist • each with a particular mouth- part structure for targeting a specific tissue type within leaves, stems, roots, and repro- ductive structures. These insects, in turn, are devoured internally by parasitoids or V- externally by parasites, each of which is attacked by hyperparasitoids, and all of which are preyed upon in complex ways by mobile and stationary insect predators. '^. . How did such intricate trophic networks originate? Is there documentation in the fos- sil record, either from the organisms themselves or from their interactions, that can illu- minate the general pattern of this increasing complexity? Can we say anything about ;/?.;-. i; species-generating processes that may be revealed by the fossil record of plant-insect associations? 'i by Conrad C. Labandeira

Geotinaes • September 1998 MACROEVOLUTIONARY STUDY OF PLANT-INSECT ASSOCIATIONS The answer to all of these questions is a qualified "yes," but so far, only prelimi- nary results have been reached. Consider a milkweed Examination and interpretation of the fossil record of plant-insect interactions plant. Within a is a new field within paleobiology, histor- ically ignored by both paleoentomolo- single photosynthesizing gists and paleobotanists and rarely dis- cussed by today's biologists studying the individual, a dozen species myriad associations between modern plants and insects. Virtually all of the of herbivorous insect existing paleobiological literature on this topic consists of descriptions of interest- consumers can co-exist. ing interactions with fossil plants, and several recent hypotheses of how insects and plants associate at macroevolution- respective host-plants, in lieu of direct ary time scales have not been addressed. fossil evidence. In this alternative Recently, phylogenetic data have been approach, molecular-based phylogenies marshaled to understand the macroevo- for plant hosts and their insect herbi- lutionary histories of selected lineages vores are established in order to deter- of extant insect herbivores and their mine if diversification occurred in paral- génies (Figure I). Rarely, however, are lel, as evidenced by the congruence of such studies explicitly calibrated by plant host and insect herbivore phylo- specimens from the fossil record; rather, they rely on a molecular clock for esti- mates of the origins of particular plant and insect taxa. Such an approach is not beetle herbivore associations hiost plant phylogeny phylogeny available for extinct lineages of asso- Hoplasoma -* *- Clerodendrum ciated plants and insects. A more direct approach is to docu- Slachyslvora -*- ment highly invariant, repeatable types of insect damage on well-preserved fos- P. quadrjmacuiata sil floras to establish distinctive patterns -^f~ S. galeficulala P. decórala •< :> of tissue use by host herbivores.

-*- S. integriilolia Temporal trends gleaned from this approach may reveal evolutionary radia- • • gf tions of certain herbivore clades onto plant taxa, or more particularly, the P. costpennis -^- trajectory of host specificity through II' time. For host specificity by insects, a P. sp. nou. -^- -^ S. mosna continuum is well known.

ÍÍ sororia -*- -*- S. drummondi

P. physostegiae -*-

S laterillora

-*^hysostegla

Figure 1. A comparison of the phylogenies of the Phyilobrotica clade of leaf Ijeeties (Chrysomelidae) and their host-plant clade of Scutellaria and related mint taxa {Lamiaceae), taken and modified fi'om Farrell and Mitter (1990). Note close correspondence between beetles and their host plants, with the notable exception of host switching of P. physostegae on Physoste^a.

September 1998 • Geotimes soft

Representing one extreme is one-to- one fidelity between an insect herbivore and its plant liost; the other extreme is rampant host-switching or collective extinction (Figure 2). A particularly intriguing method that has never been attempted in the fossil record is to track a well-established plant lineage of a few taxa through macroevo- lutionary time and monitor the spectrum of insect herbivores. Documentation of herbivore types through time can estab- lish if some insect and plant lineages are ancient associations or whether older associations are being swamped con- stantly by more recently evolved insect herbivores. If the plant host is highly con- fined taxonomically • say, a group of related species, a genus, or a nondiverse family • the fossil record can be used to examine approximately what modem biologists term a "component commu- nity." A component community constitutes all the animal associates of a particular plant-host species (Figure 3). Tracking insect herbivory within a component community, based on recognition of dam- age types, can provide data on the entries, exits, and protracted associa- tions of insect herbivore species in macroevolutionary time. Such paleobio- logical data could answer questions of high theoretical importance • such as whether co-evolved and longer-term associations occur between plants and insects, if primitive insects occur on primitive plants, or whether host-switch- ing is associated with raEyor paleoenvi- ronmental changes or with the physical features of plant hosts. :l /•:•

Figure 2. Two examples of associations iaetween leaf-mining insects and their host piants, gleaned from the fossil record of plant damage. The upper part of the figure (a) depicts an ancient •W association between an insect herbivore, the leaf-mining moth Ectoedemia, and its host plant, an undescribed form closely related to the sycamore Platsnus, from the mid-Cretaceous Dakota Formation of Kansas. The inset in the center (b) is an enlargement of this leaf mine, showing ini- tial serpentine phase and subsequent blotch phase lodged between two primary veins. The leaf on the right (c) contains near-identical Ectoedemia damage on modem Platanus from Virginia. The lower pair show either temporal host-switching or extinction of both plant host and a species of the leaf-mining moth Paraclemensia or a closely related genus. The lower-left leaf (d) exhibits Parac/emens/3-like damage on its Middle Eocene host plant Macginitiea, an extinct genus of the sycamore family from the Green River Formation of Utah. The lower-right leaf (e) features near- idenfcal Paraclemensia damage on modem Acer irom Vermont, a member of the maple family that is phylogenetlcally distant from sycamores. Scale bars: solid = 1.0 cm; striped = 0.1 cm.

Geotimes • September INCREASED INTENSITY OF INSECT HERBIVORY A second general area of paleobiological investigation involves the evaluation of insect herbivory intensity through time. Quantifying the percentage of fossil leaf area removed by insects in fossil floras has been attempted only recently, although such studies are now routine in extant tropical to warm-temperate floras. For us to understand temporal trends in insect herbivory intensity • particularly within ecologically constrained plant- conununity types • requires a test of the "Red Queen Hypothesis." This hypothesis predicts that because insects continually adapt to circumvent defenses of their plant hosts while, conversely, plant hosts develop mechanisms to ward off insect herbivory, a progressive "arms race" takes place. In other words, both herbi- vore and host are "running faster and faster just to stay in place," similar to the phght of Alice and the Red Queen in Figure 3. A component community of insect herbivores on a iiypothetical vascuiar plant. Included Through the Looking Glass. Applying are piercer-and-suchers at top (a, b) and center right (c), borers at center (d^), serpentine and this rule to the fossil record of insect her- blotch leaf miners at lower right (h), leaf and twig gailers at left

|i^^ '••m -•^l:-. September 1998 • Geotimes • \ .-. u-

'%'•• ¡m^'

(cover and Figure 4, upper-left inset), A;." sporangia and spore consumption (upper right), piercmg-and-sucking (center right), •n and galling (center left). Detritivores were present as stem and root borers After the Late Pennsylvanian, coal •'^^• (lower-right and lower-left insets, respec- swamps waned in North America. Early tively). Assignments of one mite and five Permian floras were characterized by a insect culprits to this component com- different suite of mostly seed plants, munity on Psaronius are based on diag- especially seed ferns, conifers, and :-4 This question has been investigated nostic features of the host-plant damage, cycadophytes, the latter distant relatives ;0 only recently, but the latter option coprolites with plant tissues occurring in of cycads such as the modem sago palm. appears closer to the truth. Both in terras the preserved peat deposit, and evidence However, local patches were dominated of the near-modem ecological spectrum from penecontemporaneous insect gut by gigantopterids, a group of arborescent of herbivory from a Late Pennsylvanian contents and mouthpart structures. plants with uncertain taxonomic affini- coal-swamp forest from Illinois (cover When analyzed by dietary guild (what tis- ties that bore large, sturdy, and intricate- and Figure 4) and the elevated intensity sue is consumed), Psaronius herbivores ly veined leaves. A recent examination of of live plant consumption at an Early were attacking virtually all tissue types a gigantopterid-dominated flora from the Permian riparian forest from northern • including foliage, entire sporangia or Lower Permian redbeds of northern Texas, the evidence indicates that insect individual spores, the vascular tissues of Texas revealed extensive herbivory, par- herbivory acquired a qualitative and stems and roots, and their surrounding ticularly extemal feeding in the form of quantitative head start in lovi^land equato- tissues (parenchyma). Thus, modern hole and margin consumption and skele- rial forests during the Late Paleozoic. The modes of tissue-partitioning by insects tonization of gigantopterid foUage. An currently reigning view of extensive and existed by 300 million years ago and were impressive level of herbivory • approxi- all-important Paleozoic insect detritivoiy organized into a component community mately 4 percent of the total surface area to the exclusion of significant insect her- of herbivores and detritivores. of gigantopterid leaves • was removed bivory should be abandoned. Additionally, a pattern of host-speci- by insect herbivores. This represents ficity has been documented, such as the about half the value for comparable mod- PLANT-RESOURCE USE targeting of only a certain Psaronius em tropical and subtropical angiosper- BY INSECTS organ in the cases of insects hke the peti- mous floras, Cycadophytes, fems, and The overall pattern that has emerged olar galler (35 specimens knovm) and the other plants were also consumed, but at from an examination of the 400-raillion- stem borer (seven specimens known). lesser levels. These studies indicate that year record of plant-insect associations Other, but biovolumetrically subordinate, ecologically significant•indeed elevated suggests an increase in trophic complex- plants that also were attacked by herb- • levels of herbivory were present during ity and a diversification of feeding strate- ivores include the seed-fern MeduUosa, the Early Permian. In addition, distinct gies. However, there are a few surprises. which displays extensive foUvory (leaf- host-specificity occurred in both earMer Although the earliest terrestrial ecosys- eating) of fem-like leaves, piercing-and- peat-forming wetland forests and in the tems were dominated by detritivores, sucking of bell-shaped pollen organs, subsequent drier riparian habitats domi- there is now compelling evidence that boring into stems and roots, and proba- nated by seed plants. The timing and bio- as early as the latest Silurian to Early ble consumption of individual pollen geography of expansion of insect her- Devonian, piercing-and-sucking and ex- grains that may have evolved into a bivory in a variety of distinctive plant com- ternally feeding herbivores were consum- mutualism similar to modem pollinators. munities needs to be narrowed by exam- ing live plant stems, and borers were con- Curiously, although cordaite conifers (an ining additional floras across this Late suming hyphae-bearing fungal tissues. order of gymnospermous plants that Pennsylvanian to Early Permian interval.

•/.•<•: •; Later, herbivore partitioning of plant have probably been extinct since the A dramatic extinction of most lineages of "tissues for the Late Pennsylvanian and Mesozoic) were relatively rare within the Paleozoic insect fauna occurred during Early Permian acquired a remarkably these swamp forests, most coprolites the Late Permian; the modem insect fauna, modem cast. For example, when that have spores or pollen contain characterized by the miyor insect lineages assessed by functional feeding group cordaitalean pollen. of today, emerged during the Triassic. (how insects feed), the host-plant Although the taxonomic contrasts between Psaronius • a tree fern dommant in Late these two evolutionary faunas are dramat- Pennsylvanian equatorial wetlands • ic, the feeding strategies • assessed by harbored trophically diverse herbivores mouthpart classes, functional feeding engaged in external foliage feeding groups, and dietary guilds • maintained their integrity across this boundary. These Paleozoic feeding strategies are well repre-

Geotimes • September 1998 kohle formations of France and Germany i j'V-v• •' • • indicate that all Late Paleozoic feeding ; ^y: • : sented during the Triassic, with entirely dif- strategies continued into theMesozoic. The ¡ : ferent taxa of interacting plants and earhest known occurrence of leaf-mining is mutualisms between nonangiospermous insects. Examinations of Middle and Late documented for the Late Triassic. seed plants, such as cycads, and insects TWassic plant-insect associations • such By the mid-to-late Jurassic, superbly that include weevils and flies. The sup- as those focusing on the Chinle Formation preserved biotas in Central Asia, north- port for these associations includes ^ - of the American Southwest and the Upper em China, and the western United States insect mouthpart structure, gut contents, ¿! •* Buntsandstein, Lower Keuper, and Letten- provide strong evidence for pollination damage patterns on plant reproductive organs, and certain plant reproductive structures. However, plant-insect associa- tions during the Triassic to Jurassic are poorly understood when compared to the Late Pennsylvanian to Early Permian or to the subsequent Cretaceous. During the Cretaceous, angiosperms began their extensive evolutionary and ecological radiation, providing new opportunities for partitioning vegetative resources and for elaborating mutual- isms, such as pollination, that had been established earlier with seed plants. Interestingly, the rapid extension of mid- Cretaceous angiosperms is not matched by the appearance of major new herbivo- rous feeding strategies, virtually all of which were in existence during the ear- lier Mesozoic. Additionally, the over- whelming majority of modem insect fami- lies precede the appearance of the ear- liest angiosperms in the fossil record, indicating that herbivore associations with plants • including microscale exploitation of discrete tissue types, con- sumption of broad leaves, and pollination mutualisms • occurred among seed plants in general. Evidently, these associations were subsequently elaborat- ed between insects at lower taxonomic levels and one particular (albeit currently diverse) lineage of seed plants, the angiosperms, resulting perhaps in more finely partitioned feeding strategies. These consequences are best illustrated by numerous types of external foliage feeders (including margin and hole feed- ers, skeletonizers, and surface abraders), leaf miners, gallers, and other functional feeding groups documented from the mid- Cretaceous Dakota Formation of Kansas and Nebraska. Some of these interactions Figure 4. The component community of herbivores and detritivores in their Late Pennsylvanian (300 IVla) host plant Psamnius, a tree fern that dominated peat-substrate have persisted to the present, such as phy- swamp forests of the Illinois Basin. All interactions are from insects, except the lower-right logenetically basal lineages of leaf-mining inset, which is a mite. The upper four of these associations indicate feeding of live plant tis- moths occurring on similarly basal lin- sue and are confined to the canopy. •.•;.• .'¿jA '- ' \-- eages of woody angiosperms. Pollination

September 1998 • Geotimes V-\ •i: • -- 7 A..--10 floral successions representing brief, temporally constrained intervals within macroevolutionaiy time. Such approaches syndromes became progressively elabo- provide promise by placing in appropriate rated during tlie Cretaceous and into the context the historical dimension of most Acknowledgments Cenozoic, paralleling similar fine-grained of Earth's biodiversity. Constructive reviews were provided by William DiMlciieie and Charles Mitter. I tliank Finnegan trends in plant-resource partitioning by • -.vC?' ''••"^•. Marsh for fornriatting the illustrations for this arti- herbivorous insects. During the mid- Conrad C. Labandeira ' ' ' cle and Mary Parrish for drawing the cover illus- Cenozoic, more modem lineages of Smithsonian Institution, National Museum tration. Support has been provided by the Walcott angiosperms emerged, especially grasses ofNaàiral History, NHB PaJwbiology MRC- Fund of the Smithsonian Institution. This is contri- and members of the sunflower family. 121, Washington, D.C. 20560 bution no. 38 from the Evolution oí Terrestrial Their appearance introduced a substrate Dr. Labandeira received his doctorate Ecosystems Consortium at the National Museum of Natural History. . • ... for associated insects that included from the University of Chicago, where he leafhoppers, leaf beetles, and leaf-mining completed a dissertation on a classifica- - - •-. ' -, ^' nies. This recent diversification was Additional Reading tion of modem insect mouthparts and The Coevolutionary Process by John superimposed on older Mesozoic patterns, their representation in the fossil record. Thompson. The University of Chicago Press providing the broad and diverse spectrum As a postdoctoral research fellow at the {Ctiicago), 1994. of both ancient and modem associations University of Illinois at Urbana- "Early History of Arthropod and Vascular Plant seen today. Champaign, he examined pkmt/arthro- Associations" by Conrad Labandeira. Annual pod associations in Pennsylvanian-age Review of Eartti and Planetary Sciences, v. 26, CURRENT STATUS coal-swamp forests. Since 1992, he has 1998, p. 329-377. AND FUTURE PROSPECTS been at the Smithsonian's National "The Evolution of Insect Pollination in During the past two decades, an intellec- Museum of Natural History (NMNH), Angiosperms" by William Crepet and Else Friis, in tual renaissance in the study of plant- The Origin ol Angiosperms and tiieir Biological conducting research on inject diversity Consequences, edited by E.M. Friis, W.G. Chaloner, insect interactions has taken place, hi in the fossil record and, with other col- and PR. Crane. Cambridge University Press addition to the earliest and classic leagues, examining the extent and vari- (Cambridge, U.K.), 1987, p. 181-201. escape-and-radiation model mentioned ety of herbivory in Late Pennsylvanian "Insect Diversity in the Fossil Record" by above, four other m^or hypotheses have and Early Permianfioras and Ike ecology Conrad Labandeira and J. John Sepkoski Jr. been proposed to account for observed of Cretaceous insect associates of eariy Science, v. 261, July 1993, p. 310-315. patterns of plant-host use by insects. angiosperms. They are currently working "Insect MoLthparts: Ascertaining the These alternative models involve differ- on the spectrum and intensity of insect Paleobiology of Insect Feeding Strategies" by ing temporal and spatial modes of insect herbivory from three Eocene floras. Conrad Labandeira. Annual Review of Ecology and Systematics, v. 28,1997, p. 153-193. speciation or adaptation to host plants, Labandeira is curator of fossil arthropods and vice-versa. Whether these modes of in the Department of Paleobiology at "Insect-Plart Interactions: The Evolution of Component Communities" by Douglas Futuyma plant-insect association can be resolved NMNH. and Charles Mitter. Ptiilosophical Transactions of within the temporal and spatial scales the Royal Society of London, Series B, v. 351, encountered in the fossil record is still September1996, p. 1361-1366. uncertain. In principle they should • with "A Late Carboniferous Petiole Gall and the well-preserved, diverse, and abundant Origin of Holometabolous Insects" by Conrad C. Labandeira and Tom Phillips. Proceedings of the National Academy of Sciences, USA, v. 93, August 1996, p. 8470-8474. The currently reigning "Ninety-Seven Million Years of Angiosperm- Insect Association: Paieobiological Insights into the Meaning of Coevoiution" by Conrad view of extensive and Labandeira, David Dilcher, Donald Davis, and David Wagner. Proceedings of the National all-important Paleozoic Academy of Sciences, USA, v. 91, December 1994, p. 12278-12282. insect detritivory to "Phylogenesis of InsecfPlant Interactions: Have Phyitobrotica Leaf Beetles (Chrysomelidae) and the Lamíales Diversified in Parallel?" by Brian Farrell and the exclusion of significant Charles Mitter. Evolution, v 44,1990, p. 1389-1403.

"Plant-Insect Interactions and Coevoluiion insect herbivory should be During the Triassic in Western Europe" by Lea Grauvogel-Stamm and Klaus-Peter Kelber. abandoned. Paleontológica Lombarda, v. 5,1996, p. 5-23.

Geotiraes • September 1998